The Anatomy of a Players Overall Rating

[nmm666]

Quote: Originally Posted by rudyjuly2

Those weights look very similar to last year (probably the same formula used in NCAA 12 I'm guessing). I took one of PocketScout's recruiting files and ran some regression results on them to get overall formulas for each position. Rounding is a slight issue but like the op, you get the exact rating for roughly 97 or 98% of the players. Only off by one point on the others. I'm a math geek and I love this stuff. https://spreadsheets1.google.com/spr...hl=en_US#gid=0

I've been working away doing the same thing for my offline recruiting sheet. Looks like my regression results are fairly similar to yours. I guess it varies a bit depending on the regression package and the data set being analyzed.

[JerichoGates]

Quote: Originally Posted by PocketScout

It's based on the math from my research while writing my program. Because of rounding issues 99.4% of players (497/500) were within 1 point of their actual ingame rating. So short answer it's my take on it, but a pretty accurate one overall.

That works for me! THANKS! Also are you planning on updating this with Maddens NEW 12 ratings? I don't remember exactly what they are called, but I think there are two, one is for 4th quarter performance and I can not remember off the top of my head what the other one is for.

[PocketScout]

Quote: Originally Posted by JerichoGates

That works for me! THANKS! Also are you planning on updating this with Maddens NEW 12 ratings? I don't remember exactly what they are called, but I think there are two, one is for 4th quarter performance and I can not remember off the top of my head what the other one is for.

They are all new to me since I haven't played or owned Madden in over a decade.

[nyqua]

Here's an Excel I made of the recruiting rating translations. It looks pretty nice to me.

[Th3Rush22]

Quote: Originally Posted by rudyjuly2

Those weights look very similar to last year (probably the same formula used in NCAA 12 I'm guessing). I took one of PocketScout's recruiting files and ran some regression results on them to get overall formulas for each position. Rounding is a slight issue but like the op, you get the exact rating for roughly 97 or 98% of the players. Only off by one point on the others. I'm a math geek and I love this stuff. https://spreadsheets1.google.com/spr...hl=en_US#gid=0

Can you please inform me on how to use your spreadsheets I took the players rating and multiplied it by the number on the spread sheet and than added them together but got that Matt mcgloin's overall rating was 148 so I don't think I did it right

[waytofailself]

Just wanted to thank you for the post.

edit: a little sad to see that there are less gems to be found via 3* recruiting though

[rudyjuly2]

Quote: Originally Posted by Th3Rush22

Can you please inform me on how to use your spreadsheets I took the players rating and multiplied it by the number on the spread sheet and than added them together but got that Matt mcgloin's overall rating was 148 so I don't think I did it right

Most players (other than the kickers) don't start at 0. They usually are a big negative and the starting point is indicated by the constant. For QBs the constant/starting point is -58.3. So multiple all the co-efficients by the player attributes in the formula, add them up and then also add the -58.3. See if that comes out correctly.

In the NCAA 11 database I had 95% of the guys were exactly right and the rest only off one point. I don't know if the formula will work the same way for NCAA 12. It could but I haven't checked it yet.

[Th3Rush22]

Quote: Originally Posted by rudyjuly2

Most players (other than the kickers) don't start at 0. They usually are a big negative and the starting point is indicated by the constant. For QBs the constant/starting point is -58.3. So multiple all the co-efficients by the player attributes in the formula, add them up and then also add the -58.3. See if that comes out correctly. In the NCAA 11 database I had 95% of the guys were exactly right and the rest only off one point. I don't know if the formula will work the same way for NCAA 12. It could but I haven't checked it yet.

Thanx a lot man